Deterministic and stochastic neuronal contributions to distinct synchronous CA3 network bursts

Multimodal, Multiscale Insights into Hippocampal Seizures Enabled by Transparent, Graphene-Based Microelectrode Arrays

Hippocampal seizures are a defining feature of mesial temporal lobe epilepsy (MTLE). Area CA1 of the hippocampus is commonly implicated in the generation of seizures, which may occur because of the activity of endogenous cell populations or of inputs from other regions within the hippocampal formation. Simultaneously observing activity at the cellular and network scales in vivo remains challenging. Here, we present a novel technology for simultaneous electrophysiology and multicellular calcium imaging of CA1 pyramidal cells (PCs) in mice enabled by a transparent graphene-based microelectrode array (Gr MEA). We examine PC firing at seizure onset, oscillatory coupling, and the dynamics of the seizure traveling wave as seizures evolve. Finally, we couple features derived from both modalities to predict the speed of the traveling wave using bootstrap aggregated regression trees. Analysis of the most important features in the regression trees suggests a transition among states in the evolution of hippocampal seizures.

Noise-induced network bursts and coherence in a calcium-mediated neural network

Noise-induced population bursting has been widely identified to play important roles in information processes. We construct a mathematical model for a random and sparse heterogeneous neural network where bursting can be induced from a resting state by a global stochastic stimulus. Importantly, the noise-induced bursting dynamics of this network are mediated by calcium conductance. We use two spectral measures to evaluate network coherence in the context of the network bursts, the spike trains of all neurons, and the individual bursts of all neurons. Our results show that the coherence of the network is optimized by an optimal level of the stochastic stimulus, which is known as coherence resonance (CR). We also demonstrate that the interplay of the calcium conductance and noise intensity can modify the degree of CR.

Direct Medial Entorhinal Cortex Input to Hippocampal CA3 Is Crucial for eEF2K Inhibitor-Induced Neuronal Oscillations in the Mouse Hippocampus

The hippocampal formation plays a vital role in memory formation and takes part in the control of the default neuronal network activity of the brain. It also represents an important structure to analyze drug-induced effects on subregion-specific synchronization of neuronal activity. However, the consequences of an altered functional state of synapses for subregion-specific synchronization of neuronal microcircuits remain to be fully understood. Therefore, we analyzed the direct interaction of neuronal microcircuits utilizing a genetically encoded calcium sensor (GCaMP6s) and local field potential (LFP) recording in acute hippocampal-entorhinal brain slices in response to a modulator of synaptic transmission. We observed that application of the eukaryotic elongation factor-2 kinase (eEF2K) inhibitor A484954, induced a large-scale synchronization of neuronal activity within different regions of the hippocampal formation. This effect was confirmed by the recording of extracellular LFPs. Further, in order to understand if the synchronized activity depended on interconnected hippocampal areas, we lesioned adjacent regions from each other. These experiments identified the origin of A484954-induced synchronized activity in the hippocampal CA3 subfield localized near the hilus of the dentate gyrus. Remarkably, the synchronization of neuronal activity in the hippocampus required an intact connection with the medial entorhinal cortex (MEC). In line with this observation, we detected an increase in neuronal activity in the MEC area after application of A484954. In summary, inhibition of eEF2K alters the intrinsic activity of interconnected neuronal microcircuits dominated by the MEC-CA3 afferents.

Ultra-fast accurate reconstruction of spiking activity from calcium imaging data

Calcium imaging provides an indirect observation of the underlying neural dynamics and enables the functional analysis of neuronal populations. However, the recorded fluorescence traces are temporally smeared, thus making the reconstruction of exact spiking activity challenging. Most of the established methods to tackle this issue are limited in dealing with issues such as the variability in the kinetics of fluorescence transients, fast processing of long-term data, high firing rates, and measurement noise. We propose a novel, heuristic reconstruction method to overcome these limitations. By using both synthetic and experimental data, we demonstrate the four main features of this method: 1) it accurately reconstructs both isolated spikes and within-burst spikes, and the spike count per fluorescence transient, from a given noisy fluorescence trace; 2) it performs the reconstruction of a trace extracted from 1,000,000 frames in less than 2 s; 3) it adapts to transients with different rise and decay kinetics or amplitudes, both within and across single neurons; and 4) it has only one key parameter, which we will show can be set in a nearly automatic way to an approximately optimal value. Furthermore, we demonstrate the ability of the method to effectively correct for fast and rather complex, slowly varying drifts as frequently observed in in vivo data. NEW & NOTEWORTHY Reconstruction of spiking activities from calcium imaging data remains challenging. Most of the established reconstruction methods not only have limitations in adapting to systematic variations in the data and fast processing of large amounts of data, but their results also depend on the user's experience. To overcome these limitations, we present a novel, heuristic model-free-type method that enables an ultra-fast, accurate, near-automatic reconstruction from data recorded under a wide range of experimental conditions.

Massively augmented hippocampal dentate granule cell activation accompanies epilepsy development

In a mouse model of temporal lobe epilepsy, multicellular calcium imaging revealed that disease emergence was accompanied by massive amplification in the normally sparse, afferent stimulation-induced activation of hippocampal dentate granule cells. Patch recordings demonstrated reductions in local inhibitory function within the dentate gyrus at time points where sparse activation was compromised. Mimicking changes in inhibitory synaptic function and transmembrane chloride regulation was sufficient to elicit the dentate gyrus circuit collapse evident during epilepsy development. Pharmacological blockade of outward chloride transport had no effect during epilepsy development, and significantly increased granule cell activation in both control and chronically epileptic animals. This apparent occlusion effect implicates reduction in chloride extrusion as a mechanism contributing to granule cell hyperactivation specifically during early epilepsy development. Glutamine plays a significant role in local synthesis of GABA in synapses. In epileptic mice, sparse granule cell activation could be restored by glutamine application, implicating compromised GABA synthesis. Glutamine had no effect on granule cell activation earlier, during epilepsy development. We conclude that compromised feedforward inhibition within the local circuit generates the massive dentate gyrus circuit hyperactivation evident in animals during and following epilepsy development. However, the mechanisms underlying this disinhibition diverge significantly as epilepsy progresses.

Modulation of Neural Network Activity through Single Cell Ablation: An in Vitro Model of Minimally Invasive Neurosurgery

The technological advancement of optical approaches, and the growth of their applications in neuroscience, has allowed investigations of the physio-pathology of neural networks at a single cell level. Therefore, better understanding the role of single neurons in the onset and progression of neurodegenerative conditions has resulted in a strong demand for surgical tools operating with single cell resolution. Optical systems already provide subcellular resolution to monitor and manipulate living tissues, and thus allow understanding the potentiality of surgery actuated at single cell level. In the present work, we report an in vitro experimental model of minimally invasive surgery applied on neuronal cultures expressing a genetically encoded calcium sensor. The experimental protocol entails the continuous monitoring of the network activity before and after the ablation of a single neuron, to provide a robust evaluation of the induced changes in the network activity. We report that in subpopulations of about 1000 neurons, even the ablation of a single unit produces a reduction of the overall network activity. The reported protocol represents a simple and cost effective model to study the efficacy of single-cell surgery, and it could represent a test-bed to study surgical procedures circumventing the abrupt and complete tissue removal in pathological conditions.

Inferring Neuronal Dynamics from Calcium Imaging Data Using Biophysical Models and Bayesian Inference

Calcium imaging has been used as a promising technique to monitor the dynamic activity of neuronal populations. However, the calcium trace is temporally smeared which restricts the extraction of quantities of interest such as spike trains of individual neurons. To address this issue, spike reconstruction algorithms have been introduced. One limitation of such reconstructions is that the underlying models are not informed about the biophysics of spike and burst generations. Such existing prior knowledge might be useful for constraining the possible solutions of spikes. Here we describe, in a novel Bayesian approach, how principled knowledge about neuronal dynamics can be employed to infer biophysical variables and parameters from fluorescence traces. By using both synthetic and in vitro recorded fluorescence traces, we demonstrate that the new approach is able to reconstruct different repetitive spiking and/or bursting patterns with accurate single spike resolution. Furthermore, we show that the high inference precision of the new approach is preserved even if the fluorescence trace is rather noisy or if the fluorescence transients show slow rise kinetics lasting several hundred milliseconds, and inhomogeneous rise and decay times. In addition, we discuss the use of the new approach for inferring parameter changes, e.g. due to a pharmacological intervention, as well as for inferring complex characteristics of immature neuronal circuits.

Calcium imaging of single neurons enables the indirect observation of neuronal dynamics, for example action potential firing. In contrast to the precise timing of spike trains, the calcium trace is temporally rather smeared and measured as a fluorescence trace. Consequently, several methods have been proposed to reconstruct spikes from calcium imaging data. However, a common feature of these methods is that they are not based on the biophysics of how neurons fire spikes and bursts. We propose to introduce well-established biophysical models to create a direct link between neuronal dynamics, e.g. the membrane potential, and fluorescence traces. Using both synthetic and experimental data, we show that this approach not only provides a robust and accurate spike reconstruction but also a reliable inference about the biophysically relevant parameters and variables. This enables novel ways of analyzing calcium imaging experiments in terms of the underlying biophysical quantities.

Gradient Index Microlens Implanted in Prefrontal Cortex of Mouse Does Not Affect Behavioral Test Performance over Time

Implanted gradient index lenses have extended the reach of standard multiphoton microscopy from the upper layers of the mouse cortex to the lower cortical layers and even subcortical regions. These lenses have the clarity to visualize dynamic activities, such as calcium transients, with subcellular and millisecond resolution and the stability to facilitate repeated imaging over weeks and months. In addition, behavioral tests can be used to correlate performance with observed changes in network function and structure that occur over time. Yet, this raises the questions, does an implanted microlens have an effect on behavioral tests, and if so, what is the extent of the effect? To answer these questions, we compared the performance of three groups of mice in three common behavioral tests. A gradient index lens was implanted in the prefrontal cortex of experimental mice. We compared their performance with mice that had either a cranial window or a sham surgery. Three presurgical and five postsurgical sets of behavioral tests were performed over seven weeks. Behavioral tests included rotarod, foot fault, and Morris water maze. No significant differences were found between the three groups, suggesting that microlens implantation did not affect performance. The results for the current study clear the way for combining behavioral studies with gradient index lens imaging in the prefrontal cortex, and potentially other regions of the mouse brain, to study structural, functional, and behavioral relationships in the brain.

NeuroCa: integrated framework for systematic analysis of spatiotemporal neuronal activity patterns from large-scale optical recording data

Optical recording facilitates monitoring the activity of a large neural network at the cellular scale, but the analysis and interpretation of the collected data remain challenging. Here, we present a MATLAB-based toolbox, named NeuroCa, for the automated processing and quantitative analysis of large-scale calcium imaging data. Our tool includes several computational algorithms to extract the calcium spike trains of individual neurons from the calcium imaging data in an automatic fashion. Two algorithms were developed to decompose the imaging data into the activity of individual cells and subsequently detect calcium spikes from each neuronal signal. Applying our method to dense networks in dissociated cultures, we were able to obtain the calcium spike trains of [Formula: see text] neurons in a few minutes. Further analyses using these data permitted the quantification of neuronal responses to chemical stimuli as well as functional mapping of spatiotemporal patterns in neuronal firing within the spontaneous, synchronous activity of a large network. These results demonstrate that our method not only automates time-consuming, labor-intensive tasks in the analysis of neural data obtained using optical recording techniques but also provides a systematic way to visualize and quantify the collective dynamics of a network in terms of its cellular elements.

Heterogeneous effects of antiepileptic drugs in an in vitro epilepsy model--a functional multineuron calcium imaging study

Epilepsy is a chronic brain disease characterised by recurrent seizures. Many studies of this disease have focused on local neuronal activity, such as local field potentials in the brain. In addition, several recent studies have elucidated the collective behavior of individual neurons in a neuronal network that emits epileptic activity. However, little is known about the effects of antiepileptic drugs on neuronal networks during seizure-like events (SLEs) at single-cell resolution. Using functional multineuron Ca(2+) imaging (fMCI), we monitored the activities of multiple neurons in the rat hippocampal CA1 region on treatment with the proconvulsant bicuculline under Mg(2+) -free conditions. Bicuculline induced recurrent synchronous Ca(2+) influx, and the events were correlated with SLEs. Other proconvulsants, such as 4-aminopyridine, pentetrazol, and pilocarpine, also induced synchronous Ca(2+) influx. We found that the antiepileptic drugs phenytoin, flupirtine, and ethosuximide, which have different mechanisms of action, exerted heterogeneous effects on bicuculline-induced synchronous Ca(2+) influx. Phenytoin and flupirtine significantly decreased the peak, the amount of Ca(2+) influx and the duration of synchronous events in parallel with the duration of SLEs, whereas they did not abolish the synchronous events themselves. Ethosuximide increased the duration of synchronous Ca(2+) influx and SLEs. Furthermore, the magnitude of the inhibitory effect of phenytoin on the peak synchronous Ca(2+) influx level differed according to the peak amplitude of the synchronous event in each individual cell. Evaluation of the collective behavior of individual neurons by fMCI seems to be a powerful tool for elucidating the profiles of antiepileptic drugs.

Preictal activity of subicular, CA1, and dentate gyrus principal neurons in the dorsal hippocampus before spontaneous seizures in a rat model of temporal lobe epilepsy

Previous studies suggest that spontaneous seizures in patients with temporal lobe epilepsy might be preceded by increased action potential firing of hippocampal neurons. Preictal activity is potentially important because it might provide new opportunities for predicting when a seizure is about to occur and insight into how spontaneous seizures are generated. We evaluated local field potentials and unit activity of single, putative excitatory neurons in the subiculum, CA1, CA3, and dentate gyrus of the dorsal hippocampus in epileptic pilocarpine-treated rats as they experienced spontaneous seizures. Average action potential firing rates of neurons in the subiculum, CA1, and dentate gyrus, but not CA3, increased significantly and progressively beginning 2-4 min before locally recorded spontaneous seizures. In the subiculum, CA1, and dentate gyrus, but not CA3, 41-57% of neurons displayed increased preictal activity with significant consistency across multiple seizures. Much of the increased preictal firing of neurons in the subiculum and CA1 correlated with preictal theta activity, whereas preictal firing of neurons in the dentate gyrus was independent of theta. In addition, some CA1 and dentate gyrus neurons displayed reduced firing rates preictally. These results reveal that different hippocampal subregions exhibit differences in the extent and potential underlying mechanisms of preictal activity. The finding of robust and significantly consistent preictal activity of subicular, CA1, and dentate neurons in the dorsal hippocampus, despite the likelihood that many seizures initiated in other brain regions, suggests the existence of a broader neuronal network whose activity changes minutes before spontaneous seizures initiate.

Transparent and flexible low noise graphene electrodes for simultaneous electrophysiology and neuroimaging

Calcium imaging is a versatile experimental approach capable of resolving single neurons with single-cell spatial resolution in the brain. Electrophysiological recordings provide high temporal, but limited spatial resolution, because of the geometrical inaccessibility of the brain. An approach that integrates the advantages of both techniques could provide new insights into functions of neural circuits. Here, we report a transparent, flexible neural electrode technology based on graphene, which enables simultaneous optical imaging and electrophysiological recording. We demonstrate that hippocampal slices can be imaged through transparent graphene electrodes by both confocal and two-photon microscopy without causing any light-induced artefacts in the electrical recordings. Graphene electrodes record high-frequency bursting activity and slow synaptic potentials that are hard to resolve by multicellular calcium imaging. This transparent electrode technology may pave the way for high spatio-temporal resolution electro-optic mapping of the dynamic neuronal activity.

Extracellular calcium controls the expression of two different forms of ripple-like hippocampal oscillations

Hippocampal high-frequency oscillations (HFOs) are prominent in physiological and pathological conditions. During physiological ripples (100-200 Hz), few pyramidal cells fire together coordinated by rhythmic inhibitory potentials. In the epileptic hippocampus, fast ripples (>200 Hz) reflect population spikes (PSs) from clusters of bursting cells, but HFOs in the ripple and the fast ripple range are vastly intermixed. What is the meaning of this frequency range? What determines the expression of different HFOs? Here, we used different concentrations of Ca(2+) in a physiological range (1-3 mM) to record local field potentials and single cells in hippocampal slices from normal rats. Surprisingly, we found that this sole manipulation results in the emergence of two forms of HFOs reminiscent of ripples and fast ripples recorded in vivo from normal and epileptic rats, respectively. We scrutinized the cellular correlates and mechanisms underlying the emergence of these two forms of HFOs by combining multisite, single-cell and paired-cell recordings in slices prepared from a rat reporter line that facilitates identification of GABAergic cells. We found a major effect of extracellular Ca(2+) in modulating intrinsic excitability and disynaptic inhibition, two critical factors shaping network dynamics. Moreover, locally modulating the extracellular Ca(2+) concentration in an in vivo environment had a similar effect on disynaptic inhibition, pyramidal cell excitability, and ripple dynamics. Therefore, the HFO frequency band reflects a range of firing dynamics of hippocampal networks.

Network models predict that reduced excitatory fluctuations can give rise to hippocampal network hyper-excitability in MeCP2-null mice

Rett syndrome is a severe pediatric neurological disorder caused by loss of function mutations within the gene encoding methyl CpG-binding protein 2 (MeCP2). Although MeCP2 is expressed near ubiquitously, the primary pathophysiology of Rett syndrome stems from impairments of nervous system function. One alteration within different regions of the MeCP2-deficient brain is the presence of hyper-excitable network responses. In the hippocampus, such responses exist despite there being an overall decrease in spontaneous excitatory drive within the network. In this study, we generated and used mathematical, neuronal network models to resolve this apparent paradox. We did this by taking advantage of previous mathematical modelling insights that indicated that decreased excitatory fluctuations, but not mean excitatory drive, more critically explain observed changes in hippocampal network oscillations from MeCP2-null mouse slices. Importantly, reduced excitatory fluctuations could also bring about hyper-excitable responses in our network models. Therefore, these results indicate that diminished excitatory fluctuations may be responsible for the hyper-excitable state of MeCP2-deficient hippocampal circuitry.

Synconset waves and chains: spiking onsets in synchronous populations predict and are predicted by network structure

Synfire waves are propagating spike packets in synfire chains, which are feedforward chains embedded in random networks. Although synfire waves have proved to be effective quantification for network activity with clear relations to network structure, their utilities are largely limited to feedforward networks with low background activity. To overcome these shortcomings, we describe a novel generalisation of synfire waves, and define ‘synconset wave’ as a cascade of first spikes within a synchronisation event. Synconset waves would occur in ‘synconset chains’, which are feedforward chains embedded in possibly heavily recurrent networks with heavy background activity. We probed the utility of synconset waves using simulation of single compartment neuron network models with biophysically realistic conductances, and demonstrated that the spread of synconset waves directly follows from the network connectivity matrix and is modulated by top-down inputs and the resultant oscillations. Such synconset profiles lend intuitive insights into network organisation in terms of connection probabilities between various network regions rather than an adjacency matrix. To test this intuition, we develop a Bayesian likelihood function that quantifies the probability that an observed synfire wave was caused by a given network. Further, we demonstrate it's utility in the inverse problem of identifying the network that caused a given synfire wave. This method was effective even in highly subsampled networks where only a small subset of neurons were accessible, thus showing it's utility in experimental estimation of connectomes in real neuronal-networks. Together, we propose synconset chains/waves as an effective framework for understanding the impact of network structure on function, and as a step towards developing physiology-driven network identification methods. Finally, as synconset chains extend the utilities of synfire chains to arbitrary networks, we suggest utilities of our framework to several aspects of network physiology including cell assemblies, population codes, and oscillatory synchrony.

Protracted postnatal development of sparse, specific dentate granule cell activation in the mouse hippocampus

The dentate gyrus (DG) is a critical entry point regulating function of the hippocampus. Integral to this role are the sparse, selective activation characteristics of the principal cells of the DG, dentate granule cells (DGCs). This sparse activation is important both in cognitive processing and in regulation of pathological activity in disease states. Using a novel, combined dynamic imaging approach capable of resolving sequentially both synaptic potentials and action potential firing in large populations of DGCs, we characterized the postnatal development of firing properties of DG neurons in response to afferent activation in mouse hippocampal-entorhinal cortical slices. During postnatal development, there was a protracted, progressive sparsification of responses, accompanied by increased temporal precision of activation. Both of these phenomena were primarily mediated by changes in local circuit inhibition, and not by alterations in afferent innervation of DGCs because GABA(A) antagonists normalized developmental differences. There was significant θ and γ frequency-dependent synaptic recruitment of DGC activation in adult, but not developing, animals. Finally, we found that the decision to fire or not fire by individual DGCs was robust and repeatable at all stages of development. The protracted postnatal development of sparse, selective firing properties, increased temporal precision and frequency dependence of activation, and the fidelity with which the decision to fire is made are all fundamental circuit determinants of DGC excitation, critical in both normal and pathological function of the DG.

In vitro large-scale experimental and theoretical studies for the realization of bi-directional brain-prostheses

Brain-machine interfaces (BMI) were born to control “actions from thoughts” in order to recover motor capability of patients with impaired functional connectivity between the central and peripheral nervous system. The final goal of our studies is the development of a new proof-of-concept BMI—a neuromorphic chip for brain repair—to reproduce the functional organization of a damaged part of the central nervous system. To reach this ambitious goal, we implemented a multidisciplinary “bottom-up” approach in which in vitro networks are the paradigm for the development of an in silico model to be incorporated into a neuromorphic device. In this paper we present the overall strategy and focus on the different building blocks of our studies: (i) the experimental characterization and modeling of “finite size networks” which represent the smallest and most general self-organized circuits capable of generating spontaneous collective dynamics; (ii) the induction of lesions in neuronal networks and the whole brain preparation with special attention on the impact on the functional organization of the circuits; (iii) the first production of a neuromorphic chip able to implement a real-time model of neuronal networks. A dynamical characterization of the finite size circuits with single cell resolution is provided. A neural network model based on Izhikevich neurons was able to replicate the experimental observations. Changes in the dynamics of the neuronal circuits induced by optical and ischemic lesions are presented respectively for in vitro neuronal networks and for a whole brain preparation. Finally the implementation of a neuromorphic chip reproducing the network dynamics in quasi-real time (10 ns precision) is presented.

Spatially clustered neuronal assemblies comprise the microstructure of synchrony in chronically epileptic networks

Epilepsy is characterized by recurrent synchronizations of neuronal activity, which are both a cardinal clinical symptom and a debilitating phenomenon. Although the temporal dynamics of epileptiform synchronizations are well described at the macroscopic level using electrophysiological approaches, less is known about how spatially distributed microcircuits contribute to these events. It is important to understand the relationship between micro and macro network activity because the various mechanisms proposed to underlie the generation of such pathological dynamics are united by the assumption that epileptic activity is recurrent and hypersynchronous across multiple scales. However, quantitative analyses of epileptiform spatial dynamics with cellular resolution have been hampered by the difficulty of simultaneously recording from multiple neurons in lesioned, adult brain tissue. We have overcome this experimental limitation and used two-photon calcium imaging in combination with a functional clustering algorithm to uncover the functional network structure of the chronically epileptic dentate gyrus in the mouse pilocarpine model of temporal lobe epilepsy. We show that, under hyperexcitable conditions, slices from the epileptic dentate gyrus display recurrent interictal-like network events with a high diversity in the activity patterns of individual neurons. Analysis reveals that multiple functional clusters of spatially localized neurons comprise epileptic networks, and that network events are composed of the coactivation of variable subsets of these clusters, which show little repetition between events. Thus, these interictal-like recurrent macroscopic events are not necessarily recurrent when viewed at the microcircuit scale and instead display a patterned but variable structure.

Refuting the challenges of the developmental shift of polarity of GABA actions: GABA more exciting than ever!

During brain development, there is a progressive reduction of intracellular chloride associated with a shift in GABA polarity: GABA depolarizes and occasionally excites immature neurons, subsequently hyperpolarizing them at later stages of development. This sequence, which has been observed in a wide range of animal species, brain structures and preparations, is thought to play an important role in activity-dependent formation and modulation of functional circuits. This sequence has also been considerably reinforced recently with new data pointing to an evolutionary preserved rule. In a recent “Hypothesis and Theory Article,” the excitatory action of GABA in early brain development is suggested to be “an experimental artefact” (Bregestovski and Bernard, 2012). The authors suggest that the excitatory action of GABA is due to an inadequate/insufficient energy supply in glucose-perfused slices and/or to the damage produced by the slicing procedure. However, these observations have been repeatedly contradicted by many groups and are inconsistent with a large body of evidence including the fact that the developmental shift is neither restricted to slices nor to rodents. We summarize the overwhelming evidence in support of both excitatory GABA during development, and the implications this has in developmental neurobiology.

Model-free reconstruction of excitatory neuronal connectivity from calcium imaging signals

A systematic assessment of global neural network connectivity through direct electrophysiological assays has remained technically infeasible, even in simpler systems like dissociated neuronal cultures. We introduce an improved algorithmic approach based on Transfer Entropy to reconstruct structural connectivity from network activity monitored through calcium imaging. We focus in this study on the inference of excitatory synaptic links. Based on information theory, our method requires no prior assumptions on the statistics of neuronal firing and neuronal connections. The performance of our algorithm is benchmarked on surrogate time series of calcium fluorescence generated by the simulated dynamics of a network with known ground-truth topology. We find that the functional network topology revealed by Transfer Entropy depends qualitatively on the time-dependent dynamic state of the network (bursting or non-bursting). Thus by conditioning with respect to the global mean activity, we improve the performance of our method. This allows us to focus the analysis to specific dynamical regimes of the network in which the inferred functional connectivity is shaped by monosynaptic excitatory connections, rather than by collective synchrony. Our method can discriminate between actual causal influences between neurons and spurious non-causal correlations due to light scattering artifacts, which inherently affect the quality of fluorescence imaging. Compared to other reconstruction strategies such as cross-correlation or Granger Causality methods, our method based on improved Transfer Entropy is remarkably more accurate. In particular, it provides a good estimation of the excitatory network clustering coefficient, allowing for discrimination between weakly and strongly clustered topologies. Finally, we demonstrate the applicability of our method to analyses of real recordings of in vitro disinhibited cortical cultures where we suggest that excitatory connections are characterized by an elevated level of clustering compared to a random graph (although not extreme) and can be markedly non-local.

Unraveling the general organizing principles of connectivity in neural circuits is a crucial step towards understanding brain function. However, even the simpler task of assessing the global excitatory connectivity of a culture in vitro, where neurons form self-organized networks in absence of external stimuli, remains challenging. Neuronal cultures undergo spontaneous switching between episodes of synchronous bursting and quieter inter-burst periods. We introduce here a novel algorithm which aims at inferring the connectivity of neuronal cultures from calcium fluorescence recordings of their network dynamics. To achieve this goal, we develop a suitable generalization of Transfer Entropy, an information-theoretic measure of causal influences between time series. Unlike previous algorithmic approaches to reconstruction, Transfer Entropy is data-driven and does not rely on specific assumptions about neuronal firing statistics or network topology. We generate simulated calcium signals from networks with controlled ground-truth topology and purely excitatory interactions and show that, by restricting the analysis to inter-bursts periods, Transfer Entropy robustly achieves a good reconstruction performance for disparate network connectivities. Finally, we apply our method to real data and find evidence of non-random features in cultured networks, such as the existence of highly connected hub excitatory neurons and of an elevated (but not extreme) level of clustering.